Host response-based diagnostics for identifying bacterial versus viral causes of lower respiratory infection in resource-limited settings

点击翻译按钮获取中文摘要

Project Summary/ Abstract Lower respiratory tract infection (LRTI) is a common reason for antibacterial use and misuse globally. Limitations associated with current LRTI diagnostics are a major driver of antibacterial overuse. Pathogen- based diagnostics have limited sensitivity and do not distinguish infection from colonization. In low- or middle- income countries (LMICs), LRTI diagnosis is further hindered by limited laboratory infrastructure. Host-based diagnostics that leverage the host’s response to infection and broadly classify infection as viral or bacterial in etiology could greatly reduce inappropriate antibacterial use for LRTI. Previously, we showed that novel, peripheral blood-based gene expression classifiers accurately identified bacterial versus viral febrile respiratory illness in a South Asian population. While promising, these classifiers require the collection of a blood sample, which may be challenging in pediatric populations or in LMIC settings with limited resources. Emerging data suggest that the host response in the nasopharynx may also help identify class of infection. Nasopharyngeal sampling offers the possibility of an integrated diagnostic that combines both pathogen and host response detection in a single sample, which would be especially attractive in LMIC settings. The objective of this application is to determine the performance characteristics of NP-based gene expression classifiers at differentiating viral versus bacterial LRTI in a South Asian population. The following aims are proposed 1) to derive NP-based gene expression classifiers to discriminate viral versus bacterial LRTI, and 2) to transfer the NP-based classifier to a real-time polymerase chain reaction (RT-PCR) assay that has potential to be translated to a clinical platform. Comprehensive microbiological and molecular testing for respiratory viral and bacterial pathogens will be completed. Subjects will be adjudicated as having viral versus bacterial LRTI, and RNA sequencing will be performed using NP samples. Machine-learning approaches will identify host gene expression classifiers that discriminate viral versus bacterial LRTI. The genes identified in the NP-based classifier will be migrated onto customized, TaqMan Low-Density Array (TLDA) cards and RT-PCR will be performed. Gene expression will be quantified and logistic regression performed to identify viral versus bacterial LRTI. The expected outcome of this proposal is a significant improvement in our knowledge of how novel NP-based gene expression classifiers perform at identifying viral versus bacterial LRTI in a South Asian population. Following successful completion of these aims, we plan to translate the NP-based classifier to a point-of-care, clinical diagnostic platform. The long-term goal of this work is to develop strategies for improving antibacterial use in LMICs and to help combat the global crisis of antimicrobial resistance.